Compression of Acoustic Event Detection Models With Quantized Distillation
This work addresses the deployment challenge for AED models in intelligent systems with limited computational resources, representing an incremental improvement in model compression techniques.
The paper tackles the problem of deploying large deep neural networks for Acoustic Event Detection (AED) on resource-constrained devices by proposing a compression approach that combines knowledge distillation and quantization. The result is a 15% reduction in error rate for the compact network and a model size reduction to 2% of the teacher model and 12% of the full-precision student model.
Acoustic Event Detection (AED), aiming at detecting categories of events based on audio signals, has found application in many intelligent systems. Recently deep neural network significantly advances this field and reduces detection errors to a large scale. However how to efficiently execute deep models in AED has received much less attention. Meanwhile state-of-the-art AED models are based on large deep models, which are computational demanding and challenging to deploy on devices with constrained computational resources. In this paper, we present a simple yet effective compression approach which jointly leverages knowledge distillation and quantization to compress larger network (teacher model) into compact network (student model). Experimental results show proposed technique not only lowers error rate of original compact network by 15% through distillation but also further reduces its model size to a large extent (2% of teacher, 12% of full-precision student) through quantization.